• If You Are Having Trouble Logging In with Your Old Username and Password, Please use this Forgot Your Password link to get re-established.
  • Hey! Be sure to login or register!

Weld failures

fancypants

FOUNDER
IL
There's an article in this month's EAA Experimenter that may be of interest to those that are scratch building their planes. It's a trimmed down version of the NTSB factual report for the 2011 Wright Model B accident with some good pictures of the bad welds. The report does not specify what process was used, so I'm assuming the parts were TIG welded. One news article states that the work was contracted out but did not name the welder.

Abstract here
Full narrative here
PDF of the EAA Experimenter article attached View attachment EAA Experimenter - March 2013 - Wright Flyer Weld Failure.pdf

I'm not building anything yet and have only received basic training on oxy-acetylene welding. I'm only bringing this up to satisfy my curiosity. A few questions come to mind:

  • What non-destructive methods can be used to inspect welds in 4130 tubing?
  • Is there anything about the weld shown in figure 5 (aside from the obvious fracture) that might indicate it would be prone to failure?
  • Assuming that a competent welder is doing the work, is any process more likely to have complete penetration of the work (OA vs. TIG)?

It goes without saying that a propeller drive shaft will be under different stresses than a cluster of 4130 fuselage tubing, but I figure there must be something to be learned from this fatal accident.


fig01.jpg
Figure 1—The two fractured propeller shafts, as received. The left side shaft had completely fractured at the forward weld, while the right side shaft was fractured but still attached in the same location.


fig02.jpg
Figure 2—The aft fracture surface of the failed forward weld on the left side propeller shaft tube, as received.


fig03.jpg
Figure 3—The forward fracture surface of the failed forward weld on the left side propeller shaft tube, as received. This is the mating side of Figure 2.


fig04.jpg
Figure 4—The fractured forward weld of the right side propeller shafttube assembly, showing (a) a side with complete fracture and (b) the small section still intact.


fig05.jpg
Figure 5—The aft weld of the left propeller shaft, as received, showing a visible crack along the weld line.


fig06.jpg
Figure 6—Typical section of the forward side fracture surface of the left side propeller shaft forward weld. The areas of fracture, lack of weld penetration, and smearing damage are labeled. Pores were also present on the fractured areas (~50X magnification)


fig07.jpg
Figure 7—Typical section of the aft side fracture surface of the right side propeller shaft forward weld, after cleaning. The areas of fracture, lack of weld penetration, and smearing damage are labeled. Pores were also present in the fractured areas of this weld (~30X magnification)


fig08.jpg
Figure 8—Closer view of a typical area of the right side propeller shaft forward weld fracture surface, after cleaning. An area of fatigue could be observed between the unwelded inner area and the rest of the fracture surface.


fig09.jpg
Figure 9—Secondary electron (SE) micrograph of the left propeller shaft fracture surface, showing fatigue striations near a pore defect, after cleaning. The boxed area is shown in Figure 10.


fig10.jpg
Figure 10—SE micrograph of the boxed area in Figure 9, showing fatigue striations.


fig11.jpg
Figure 11—Secondary electron (SE) micrograph of the right propeller shaft fracture surface, showing fatigue striations that developed at the unwelded areas of the forward weld, after cleaning.


fig12.jpg
Figure 12—SE micrograph of the right propeller shaft fracture surface, showing areas of overstress and post-fracture surface contamination.


fig13.jpg
Figure 13—Cross-section of an intact section the right side propeller shaft aft weld, showing the depth of the weld penetration relative to the joint (~25X, etched with 4% Nital).


fig14.jpg
Figure 14—Cross-section of an cracked section the right side propeller shaft aft weld, showing a crack emanating from a gap that had not been welded (~25X, etched with 4% Nital).
 

Attachments

  • EAA Experimenter - March 2013 - Wright Flyer Weld Failure.pdf
    2.1 MB · Views: 267
  • fig01.jpg
    fig01.jpg
    110.4 KB · Views: 1,342
  • fig02.jpg
    fig02.jpg
    101.4 KB · Views: 1,250
  • fig03.jpg
    fig03.jpg
    88.2 KB · Views: 1,224
  • fig04.jpg
    fig04.jpg
    129.6 KB · Views: 1,225
  • fig05.jpg
    fig05.jpg
    66.4 KB · Views: 1,237
  • fig06.jpg
    fig06.jpg
    152.2 KB · Views: 1,213
  • fig07.jpg
    fig07.jpg
    86.8 KB · Views: 1,207
  • fig08.jpg
    fig08.jpg
    244.2 KB · Views: 1,203
  • fig09.jpg
    fig09.jpg
    192.1 KB · Views: 1,204
  • fig10.jpg
    fig10.jpg
    176.4 KB · Views: 1,185
  • fig11.jpg
    fig11.jpg
    162.7 KB · Views: 1,196
  • fig12.jpg
    fig12.jpg
    126.8 KB · Views: 1,236
  • fig13.jpg
    fig13.jpg
    160.9 KB · Views: 1,235
  • fig14.jpg
    fig14.jpg
    179.4 KB · Views: 1,226
To me it looks like the filler was the only material holding the pieces together (with no substrate penetration), and quite a wide section too. Also appears to be TIG, which some argue for post weld heat treatment. Not sure if factor here.

I wouldn't let this discourage you from scratch building an airplane. I think it is a very rare occasion of a weld failure in any homebuilt. Hundreds have been built over the years. Take your time to learn with scrap 4130, when satisfied, start with rudder pedals, then work your way from the tail to the engine mount. By the time you get to the engine mount you are a pro! ;)
 
Last edited:
Cubs were engineered for an average gas welder to assemble. I have seen some cold welding by beginners that should be scrapped or grinded out and re-welded. Most homebuilders I have seen are honorable and would not let any substandard assembly occur on their project. But, we can't be everywhere all the time to see all fabrication processes. Therefore I am on high alert when checking out "tube and fabric" even on a factory new bird. Humans will make mistakes, good Q/C and periodic follow ups are paramount!

My good old Aztec had gone through hell several times here in Hawaii and she never cracked a weld. The front gear cage busted out of a friend's PA23-250F because a student landed directly on the front gear after a porpoise. It was a one inch crack that cost (2) overhaul engines and props, plus a crane to get her off the runway, about $100K (2007 Dollars).

A cop, long ago, told me to trust no one! I took it further to include myself.
 
The weld is not solely to blame, it is the design.
This should have been a shaft with splines on each end in which the bearing races press on and are retained with a snap ring or some other method. Also 4130 is crap, at a minimum it should be 4340 or 300M steel.

Any shaft that will receive the majority of its load through torque needs to have the weld area lengthened as in the picture below, also Rosette welds would have been beneficial.

Assem1.JPG
 

Attachments

  • Assem1.JPG
    Assem1.JPG
    201.6 KB · Views: 189
The weld is not solely to blame, it is the design.
This should have been a shaft with splines on each end in which the bearing races press on and are retained with a snap ring or some other method. Also 4130 is crap, at a minimum it should be 4340 or 300M steel.

Any shaft that will receive the majority of its load through torque needs to have the weld area lengthened as in the picture below, also Rosette welds would have been beneficial.

That is what the article should say in the first paragraph! Good common sense was the issue (opinion).
 
i agree with all the above statements. looks to me like there was not enough penetration nor proper prep before welding.ie beveling etc. What rod was used? Who welded it? sorry buggs i don't agree. Just because you welded up a few rudder pedals and some tubing on one airplane does not make you a pro. even with proper instruction it takes years to really know and understand welding techniques and procedures. I understand there are people out there that pick up on stuff pretty quick but NOT everyone. i have seen some very terrible welds from people that though they were pretty good. As i said in another post. just because you can stick two pieces of metal together does not make you a welder!

John
 
I didn't know squat about welding until I built my plane. I took a course a community college on stick welding. Then taught myself to O/A weld. After this project I have might not be an underwater welder but I consider myself a good welder. Sure, the "pro" remark is tongue in cheek but everyone has a chance to be good at it. Practice first.
 
Last edited:
you're right Buggs, everyone has a chance at it. My point was to get some sort of training, like you did and then practice.Don't just go welding on your own aand learn a lot of bad habits. Like I said there are people out there that catch on pretty quick. A lot of those early tubing experimentals were built as a joint project. Somebody knew somebody that was a welder or theylearned from somebody how to stick two pieces of metal together. I've looked at a lot of those airplanes and some of them were pretty good and others were pure garbage. get some training from a school or someone that is a welder is all i meant. didn't mean to ruffle any feathers.

John
 
Look at the size of that bearing and shaft. Then tubing that has to hold it all together. It must be perfectly alined with no play at all. After Tig welded ,if let to cool on its own it will not be hard, but soft around weld. The thickness of tubing to me does not fit the bearing.
Mark
 
Looking at fig.4 and 5 the weld itself failed. From these pictures it looks like there was very little penetration. Proper attention to procedure for welding that kind of joint could have prevented that kind of failure. Although i do agree with Mike about the vibration contributing to the failure. If it was welded properly the failure would have likely occurred on the edge of the weld not down the middle. But a failure is a failure.

John
 
The question still unanswered in my mind relates to visual inspection. Is there anything about the appearance of the welds that would suggest inadequate penetration?

In figure 5 below, you can see the weld was started at the red line, continued "up" and around to end at the red line. The first half of the weld looks to be undercut slightly (black arrow). Too little filler? Too low of an amperage? Or does it look like a passable weld?

I guess I'd like to know if a well trained eye would look at this weld and say "nope, do it again".

fig05 edited.jpeg
 

Attachments

  • fig05 edited.jpeg
    fig05 edited.jpeg
    137.9 KB · Views: 170
The black line looks like the guy burned out the base metal and did not properly refill with rod. I would scarp or re weld if permitting. No real bead.
 
The question still unanswered in my mind relates to visual inspection. Is there anything about the appearance of the welds that would suggest inadequate penetration?

In figure 5 below, you can see the weld was started at the red line, continued "up" and around to end at the red line. The first half of the weld looks to be undercut slightly (black arrow). Too little filler? Too low of an amperage? Or does it look like a passable weld?

I guess I'd like to know if a well trained eye would look at this weld and say "nope, do it again".

View attachment 10600
I'd say the weld "looks" superficial (shallow), laid down with moderate heat and high travel speed, but that's a pretty off-hand perspective. If it was laid down fast, as it looks like it might be, there could also be some self-quenching and embrittlement involved.
 
That's what I said, too little filler. It does look like the parent material was thinned into the weld.. The bulb on a weld adds to its strength, at least that's what the FAA books say I've read.

Now, I wouldn't weld this shaft or fly a welded prop shaft if it were the best welder in the world. Mike's right, way too much pulsing, vibration, uneven air loads, precession, torque, and all that. Concentrically balanced? Perfectly round? If you look it up, most engines that have drive boxes, belt reducers, prop extensions, and basically anything other than a crankshaft to a prop have issues.

Mongo's right too. Rosette, tube in a tube, and cuff it. I still wouldn't do it.

If you use 4130 filler on 4130 parent, you really need to heat treat. Most frames, if not nearly all frames are 4130 tubing with mild steel filler which meets strength requirements easily, but doesn't have embrittlement issues. Normal air cooling "normalizing" works fine.

Welds can be X-rayed. I don't know much about it or cost.

Im an amatuer welder who reads and took a class like buggs, take advice accordingly.

Weld frame - fine
Weld a prop shaft - are you f nuts??!!
 
Just curious, was this shaft an exact copy of the one that the Wright brothers put on their plane? They managed to fly their plane quite a bit without this type of failure.
 
They managed to fly their plane quite a bit without this type of failure.

Compared to what? How many times did they land in unintended fields due to engine/prop failure?

Reliability of aircraft has increased over the years... some of the old ways are best left behind.
 
Compared to what? How many times did they land in unintended fields due to engine/prop failure?
They did not die from a prop failure. That tells me that the props stayed on the plane, therefor the subject prop shafts did not fail. Had their props departed it is likely that the airplane would have received some serious damage causing loss of control.

Reliability of aircraft has increased over the years... some of the old ways are best left behind.
Agreed, perhaps the Wrights prop shafts were of different design than those on this replica?
 
The weld is not solely to blame, it is the design.
This should have been a shaft with splines on each end in which the bearing races press on and are retained with a snap ring or some other method. Also 4130 is crap, at a minimum it should be 4340 or 300M steel.

Any shaft that will receive the majority of its load through torque needs to have the weld area lengthened as in the picture below, also Rosette welds would have been beneficial.

View attachment 10577


I agree with Mongo on this. It is simply a bad design. I don't think it would have mattered who welded it together or how perfect the weld was. The design was set up for failure. In this case the weld failed. Had the weld not failed the tubing outside of the weld would have. The part that sucks it two people lost their lives because of it. The point here is look at all the experimental Super Cubs on this site. I love experimentals and the fact that the Feds allow these to be built. It has given these airplanes the ability to take off and land where very few others can, increase gross weight, etc. However, this can be the price you pay to experiment. Bottom line is test the mods you use and continue a close inspection on anything you change. My guess is that this failure was in the making and if inspected closely the cracks may have been found before the catastrophic failure occured. Just keep a close eye on everything, especially things that have been modified. We all get into bad habits of doing a less than stellar preflights.......me included. Look that airplane over good, no matter what kind it is. Super Cubs take a lot of abuse so they probably need a little more attention than most. When's the last time we looked the strut forks over. I mean clean them and really look them over. Higher gross weights mean more stress and the forks have been a problem for Piper airplanes. So we replace tham and forget them. Once in a while guys we need to spend some time ourselves and really look our aircraft over. Might be surprised at what we find too!

Steve
 
I am no expert on the matter, but I vote +1 on Mongo´s comment. Stress will build on the weld are regardless of the welding technique applied, and probably would fail first on the welded area first a Steve mentions (C-90 Cub.)
I think that shear and torsion buildup as depicted in Fig 9 is a clear indication that this would have happen even on the material itself should the weld held up as fatigue was already evident in the material itself. It is hard to know where that fatigue was coming from either under specked material as part of the design failure or if there was something else that lead to the fatigue...
 
To answer your questions, nondestructive testing on a Cub or home built fuselage is not necessary. If it looks good it will be good!
The problem with the shaft is not the weld, it is the design of the joint. This design works great for non torsion loads. Like an "A" arm for a race car or Baja rig! What gets me is there isn't even the minimum inner tube length as if this were a repair. If this were a tube repair on a fuselage it would require at least 2.5 D. Like others have stated a fish mouth or some rosettes would have helped distributed the load over a greater area instead of localizing on the weld. I would bet this was not trued in a lathe after it was welded. If it was the best TIG weld that could have been made this joint still would have failed right at the weld instead of in the center of the weld. It appears there is a miss alignment in figure 5 on the top side the bottom appears true. Knowing this, every time the shaft rotated there would be a fatigue cycle. It seems like I remember steel could handle half of its PSI strength in fatigue cycle so if 4130's PSI is around 120,000 and this shaft averaged 1800rpm I would guess this shaft would last about 33 hours. This is just a guess I'm not an Engineer. If I had to weld this joint as it is designed for some other application I would have brazed it. I believe a U joint might have helped prevent this.
 
I just welded my shock struts. Took a look inside the tube and can see the outline of the weld as in 100% penetration, not just the outline of the HAZ.
 
Good post, also good review and reminder to check your work... If you aren't sure about what you are doing when welding structural or moving parts. Hire someone who does....
 
Back
Top